NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
crates.io
crates.io, #4285 most downloaded on crates.io
Arrow integration for pyo3.
Latest stable
0.21.0
Last release 18 days ago
04 Nov 2026
Ships fairly regularly
a new release about every 5 weeks
Nearly every release is documented
notes for 27 of 27 stable releases
Nothing withdrawn
no release was ever pulled
2 years old
28 releases · first in 2024
Drag across the ruler to compare, or tap a release to open it.
Release with pyo3 0.30 and arrow 60.1. It's a minor bump because #551 moves to arrow 60 and #559 replaces pyo3-arrow's ArrayReader and ArrayIterator w
Release with pyo3 0.30 and arrow 60.1. It's a minor bump because #551
moves to arrow 60 and #559 replaces pyo3-arrow's ArrayReader and
ArrayIterator with the arrow-rs types.
Co-Authored-By: Claude Opus 5.5 noreply@anthropic.com
The pyo3-arrow changelog is for crates that use pyo3-arrow to move Arrow
data between Rust and Python. Changes to the arro3 Python API that lives
in pyo3-arrow (#549, #568, #569, #570, #571) belong in arro3's release
notes instead.
Co-Authored-By: Claude Opus 5.5 noreply@anthropic.com
Co-authored-by: Claude Opus 5.5 noreply@anthropic.com
Release with pyo3 0.30 and arrow 60.
ArrayReader and ArrayIterator are now the arrow-rs types: import them from arrow_array instead of pyo3_arrow::ffi. ArrayReader::field returns &FieldRef (#559)get_last_error returns NULL, and its errors include the producer's error code (#559)rust-version), which arrow 60.1 requiresYour coding agent can read these notes before it upgrades. Set up the MCP server →
Co-Authored-By: Claude Opus 5.5 noreply@anthropic.com
Co-Authored-By: Claude Opus 5.5 noreply@anthropic.com
Co-authored-by: Claude Opus 5.5 noreply@anthropic.com
Release with pyo3 0.30 and arrow 59.
chore: Bump pyo3-arrow to 0.19
chore: Bump pyo3-arrow to 0.19 (#508)
pyo3-arrow is not part of the workspace, so we can't use -p , which stands for package I think
pyo3-arrow is not part of the workspace, so we can't use -p, which stands for package I think
Release with pyo3 0.28 and arrow 59.
chore: Bump pyo3-arrow 0.17
chore: Bump pyo3-arrow 0.17 (#487)
chore: Bump pyo3-arrow to 0.16.1
chore: Bump pyo3-arrow to 0.16.1 (#485)
chore: Prepare pyo3-arrow for 0.16
chore: Prepare pyo3-arrow for 0.16 (#472)
chore: Bump to arrow 57 update changelog
chore: Bump to arrow 57
update changelog
Release with pyo3 0.26 and arrow 57.
Release with pyo3 0.26 and arrow 57.
Release with pyo3 0.27 and arrow 56.
Release with pyo3 0.27 and arrow 56.
Rename to_pyarrow to into_pyarrow because it consumes self. #391
to_pyarrow to into_pyarrow because it consumes self. #391export::Arro3ArrayReader by @kylebarron in https://github.com/kylebarron/arro3/pull/354Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.11.0...pyo3-arrow-v0.12.0
Improved array formatting of PyArray, PyChunkedArray, PyRecordBatch, PyScalar, and PyTable in the __repr__ exposed to Python if you re-export these ty
arrow 56.PyArray, PyChunkedArray, PyRecordBatch, PyScalar, and PyTable in the __repr__ exposed to Python if you re-export these types in your Python library #335.Fix pyo3-arrow error when importing buffer protocol object. We were incorrectly validating the stride. But since we already checked the buffer to be C
make_array now that we have a fix to https://github.com/apache/arrow-rs/issues/6151 (Part of #328).Bump pyo3 to 0.25 in pyo3-arrow by @kylebarron in https://github.com/kylebarron/arro3/pull/327
Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.9.0...pyo3-arrow-v0.10.0
feat: add batch_size parameter to read_parquet by @kevinjacobs-delfi in https://github.com/kylebarron/arro3/pull/494
Full Changelog: https://github.com/kylebarron/arro3/compare/py-v0.8.2...py-v0.9.0
arrow; instead depend on individual arrow crates for a smaller dependency tree and faster compile times.Bump pyo3-arrow to pyo3 0.24 by @kylebarron in https://github.com/kylebarron/arro3/pull/300
Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.7.2...pyo3-arrow-v0.8.0
DataType is now exported through __arrow_c_schema__ (such as when passing to pyarrow.field) with nullable: true. This is required to match equality semantics with pyarrow. by @kylebarron in https://github.com/kylebarron/arro3/pull/483Full Changelog: https://github.com/kylebarron/arro3/compare/py-v0.7.0...py-v0.8.0
Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.7.2...pyo3-arrow-v0.8.0
Add into_arro3 method https://github.com/kylebarron/arro3/pull/289
into_arro3 method https://github.com/kylebarron/arro3/pull/289Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.7.1...pyo3-arrow-v0.7.2
Bump to pyo3 0.23.4 https://github.com/kylebarron/arro3/pull/287
Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.7.0...pyo3-arrow-v0.7.1
Add frozen parameter to pyclass macro by @kylebarron in https://github.com/kylebarron/arro3/pull/282
PyObject and add wrapper types for arro3 export by @kylebarron in https://github.com/kylebarron/arro3/pull/269Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.6.0...pyo3-arrow-v0.7.0
Table.remove_column return IndexError by @surister in https://github.com/kylebarron/arro3/pull/442names to columns parameter in Tables.drop_columns by @surister in https://github.com/kylebarron/arro3/pull/453IndexError on incorrect index in Table.add_column by @surister in https://github.com/kylebarron/arro3/pull/455Table.append_column support Array by @surister in https://github.com/kylebarron/arro3/pull/433Table.set_column and Table.add_column support Array and ArrayReader by @surister in https://github.com/kylebarron/arro3/pull/437Table.drop_columns by @surister in https://github.com/kylebarron/arro3/pull/440Table.rename not modifying internals batches schema by @surister in https://github.com/kylebarron/arro3/pull/457RecordBatch.from_arrays closely to pyarrow by @AlenkaF in https://github.com/kylebarron/arro3/pull/461pyproject.toml by @kylebarron in https://github.com/kylebarron/arro3/pull/480Full Changelog: https://github.com/kylebarron/arro3/compare/py-v0.6.5...py-v0.7.0
Add test for segmentation fault when converting arro3.core.Array to pyarrow.Array at interpreter exit by @3ok in https://github.com/kylebarron/arro3/p
Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.5.1...pyo3-arrow-v0.6.0
arro3-coreNo breaking changes.
Array.from_numpy. It now additionally supports datetime64, timedelta64, fixed-width strings and bytes, variable-width strings, and object dtypes that contain str or bytes.arrow 56 by @kylebarron in https://github.com/kylebarron/arro3/pull/368. This includes all changes in the upstream Changelog.Full Changelog: https://github.com/kylebarron/arro3/compare/py-v0.5.1...py-v0.6.0
chore: Use released pyo3-file by @kylebarron in https://github.com/kylebarron/arro3/pull/341
Full Changelog: https://github.com/kylebarron/arro3/compare/py-v0.5.0...py-v0.5.1
Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.5.0...pyo3-arrow-v0.5.1
no-default-features for pyo3-arrow by @kylebarron in https://github.com/kylebarron/arro3/pull/232Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.5.0...pyo3-arrow-v0.5.1
Bump to pyo3 0.22 by @kylebarron in https://github.com/kylebarron/arro3/pull/226
Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.4.0...pyo3-arrow-v0.5.0
__repr__ by @kylebarron in https://github.com/kylebarron/arro3/pull/335DataType hashable by @kylebarron in https://github.com/kylebarron/arro3/pull/336Schema is iterable by @kylebarron in https://github.com/kylebarron/arro3/pull/338ArrowStreamExportable by @kylebarron in https://github.com/kylebarron/arro3/pull/339Full Changelog: https://github.com/kylebarron/arro3/compare/py-v0.4.6...py-v0.5.0
Zero-copy buffer protocol data import by @kylebarron in https://github.com/kylebarron/arro3/pull/204
Zero-copy buffer protocol data import by @kylebarron in https://github.com/kylebarron/arro3/pull/204
New compute functions:
Implement scalar equality by @kylebarron in https://github.com/kylebarron/arro3/pull/205
Improved docstrings by @kylebarron in https://github.com/kylebarron/arro3/pull/198
Add installation section to readme by @kylebarron in https://github.com/kylebarron/arro3/pull/189
RecordBatch.column by @kylebarron in https://github.com/kylebarron/arro3/pull/191Full Changelog: https://github.com/kylebarron/arro3/compare/py-v0.3.1...py-v0.4.0
FromPyObject for PyScalar by @kylebarron in https://github.com/kylebarron/arro3/pull/199Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.3.0...pyo3-arrow-v0.4.0
Wheels for pyodide are auto-built on CI and published to Github releases.
__getitem__ with a negative index by @kylebarron in https://github.com/kylebarron/arro3/pull/171Full Changelog: https://github.com/kylebarron/arro3/compare/py-v0.2.1...py-v0.3.0
requested_schema argument. If the cast is invalid, the original data is exported.PyScalar class for managing Arrow scalar interop.PyArray and PyScalar implement Datum.thiserror for the error enum.AnyDatum input object for allowing either array or scalar input.fix readme indentation by @kylebarron in https://github.com/kylebarron/arro3/pull/130
Full Changelog: https://github.com/kylebarron/arro3/compare/py-v0.2.0...py-v0.2.1
Full Changelog: https://github.com/kylebarron/arro3/compare/pyo3-arrow-v0.2.0...pyo3-arrow-v0.2.1
In general, high parity with pyarrow for most data access and management. Most of the Table, ChunkedArray, Array, RecordBatchReader, Schema, and Field
Table, ChunkedArray, Array, RecordBatchReader, Schema, and Field methods and behavior should be similar to pyarrow.ArrayReader, an abstraction beyond RecordBatchReader to allow a stream of arrow arrays that are not RecordBatches.ArrayReader. It parallels RecordBatchReader but is more general, supporting arbitrary Arrow arrays that do not have to represent a record batch.AnyArray enum that supports either Array or ArrayReader input.Table::new to Table::try_new and ensure that all batches have the same schema. Similar for Array::new and ChunkedArray::new.Table::newInitial arrow definitions for Array, ChunkedArray, Table, RecordBatch, Schema, and Field by @kylebarron in https://github.com/kylebarron/arro3/pull/1
compute and internal modules by @kylebarron in https://github.com/kylebarron/arro3/pull/12AnyRecordBatch: enum over data interface and stream interface. by @kylebarron in https://github.com/kylebarron/arro3/pull/25to_pyarrow method on FFI structs by @kylebarron in https://github.com/kylebarron/arro3/pull/28AsRef on PyArray and PyChunkedArray by @kylebarron in https://github.com/kylebarron/arro3/pull/33Full Changelog: https://github.com/kylebarron/arro3/commits/pyo3-arrow-v0.1.0
Nothing published for this version
One column per month.